All pull requests

Pull requests

• fix(tts): AudioPlayer.stop() drops audio that never reached the auto-start threshold

  #1001·bennyliaw opened Oct 6, 2026·2 of 3 tasks·

  ·
  1

• docs(qwen3_tts): Base checkpoints have no preset speakers

  #1000·hsliuustc0106 opened Oct 6, 2026·

  ·
  1

• fix(qwen3_tts): make the 0.6B instruct guard reachable

  #999·hsliuustc0106 opened Oct 6, 2026·

  ·
  1

• fix(server): forward prompt to STT models on /v1/audio/transcriptions

  #997·Bannng opened Oct 5, 2026·

  ·
  1

• fix(server): use Qwen3-TTS generation config for omitted sampling params

  #995·chnln opened Oct 4, 2026·3 tasks done·

  ·
  1

• Add Irodori-TTS v4-Large support (T5Gemma2 text encoder)

  #993·yoshphys opened Oct 4, 2026Contributor·2 of 3 tasks·

  ·

• Add Nemotron speech streaming STT

  #992·Lazarus-931 opened Oct 3, 2026Collaborator·

  ·
  1

• perf(breeze): keep sampled depth tokens on the device (builds on #966)

  #988·EdwardGong opened Sep 30, 2026Contributor·1 of 3 tasks·

  ·
  1

• feat(stt): expose Qwen3-ASR token log probabilities

  #985·radraccoon opened Sep 29, 2026·

  ·
  1

  2

• fix(stt): tokenizer-safe streaming detokenization

  #983·pszemraj opened Sep 29, 2026·2 of 3 tasks·

  ·

• feat(stt): add Granite Speech 4.1 Plus support

  #978·pszemraj opened Sep 27, 2026·3 tasks done·

  ·
  6

• Add streaming support for Fish S2 Pro

  #977·x-n2o opened Sep 26, 2026·3 tasks done·

  ·
  1

• Fix model category probing when optional dependencies are missing

  #974·MohammadHijjawi97 opened Sep 25, 2026·2 of 3 tasks·

  ·
  1

  1

• perf(breeze): cache the depth decoder KV per frame (RTF 1.63 -> 0.62)

  #966·jolionlands opened Sep 19, 2026·1 of 3 tasks·

  ·
  3

• fix(whisper): ignore empty non-speech token encodings

  #934·saichowdary007 opened Aug 30, 2026·

  ·
  1

  2

• fix(stt): validate empty Qwen3-ASR audio

  #931·ainergiz opened Aug 29, 2026Contributor·

  ·
  1

  1

• fix(stt): stop fabricating Qwen3-ASR stream timestamps

  #930·ainergiz opened Aug 29, 2026Contributor·

  ·
  1

  1

• fix(server): bound resident model memory with LRU eviction and idle TTL sweeper

  #913·rkfshakti opened Aug 27, 2026·3 tasks done·

  ·
  1

  13

• Onboarding: offline Dia tests + guide updates for quant predicates, KV cache sizing, and batch contracts

  #909·mm65x opened Aug 24, 2026Contributor·

  ·
  1

• fix(qwen3_tts): yield silent audio instead of dropping empty batch items

  #907·mm65x opened Aug 24, 2026Contributor·

  ·
  1

• docs(qwen3_tts): clarify voice cloning vs speaker synthesis usage rules

  #906·mm65x opened Aug 24, 2026Contributor·

  ·
  1

• fix(fish_speech): prevent fast_ path and embeddings from quantizing to fix convert script

  #905·mm65x opened Aug 24, 2026Contributor·

  ·
  1

• fix(dia): allocate kv cache to fit prefill + generation length

  #904·mm65x opened Aug 24, 2026Contributor·

  ·
  2

• fix(server): stream one container per response instead of one per chunk

  #899·arvindvenkataramani opened Aug 21, 2026·2 of 3 tasks·

  ·
  1

  2

• server: bound ModelProvider's model cache with LRU eviction

  #884·mstkyvz opened Aug 11, 2026·

  ·
  1

  1